RE: [HACKERS] MVCC works in serialized mode! - Mailing list pgsql-hackers

From Hiroshi Inoue
Subject RE: [HACKERS] MVCC works in serialized mode!
Date
Msg-id 000501be3dc4$23088d80$2801007e@cadzone.tpf.co.jp
Whole thread Raw
In response to Re: [HACKERS] MVCC works in serialized mode!  (Vadim Mikheev <vadim@krs.ru>)
List pgsql-hackers
Hello all,

> -----Original Message-----
> From: root@dune.krs.ru [mailto:root@dune.krs.ru]On Behalf Of Vadim
> Mikheev
> Sent: Monday, January 11, 1999 1:26 AM
> To: Hiroshi Inoue
> Cc: hackers@postgreSQL.org
> Subject: Re: [HACKERS] MVCC works in serialized mode!
>
>
> Hiroshi Inoue wrote:
> >
> > I have a question about new default(isolation level / transaction mode).
> >
> > How do we upgrade existent programs,if default is different
> > from current ?
> > I think the isolation level of current PostgreSQL is SERIALIZABLE
> > and the transaction mode is un-chained..
> >
[snip]
>
> 1. All our applications run in serializable mode now.
> 2. "Environment" is changed from locking to
>    multi-versioning.
>
> This has to be explained in release notes.
>
> Should we implement ability to run backend in mode compatible
> with old versions (it seems easy to do - just use AccessExclusive
> Lock for UPDATE/INSERT/DELETE in Executor and don't release
> AccessShare Lock in heap_endscan) ?
>

Yes,if it's easy as you say.
But I don't know whether other people mind it or not.
If no one mind it,SERIALIZABLE is permissible for me.

BTW before user's code of our own,there are no problems with
interface library such as ODBC/JDBC etc ?
Those work well with chained mode or read committed isolation
level ?

Thanks.

Hiroshi Inoue
Inoue@tpf.co.jp



pgsql-hackers by date:

Previous
From: "Jackson, DeJuan"
Date:
Subject: CONSTRAINTS...
Next
From: "Thomas G. Lockhart"
Date:
Subject: Re: [HACKERS] postgres and year 2000